Disclosure: this article contains affiliate links. If you buy through them we may earn a commission at no additional cost to you. Our recommendations are based on the evidence recorded alongside this article, and prices were verified on the date shown.

What changed and why it matters

Science.org, the publisher behind Science magazine, ran a blog post this month with a title that reads more like a challenge than a headline: "So How Is AI Drug Discovery Doing, Really?" (read it directly). That a mainstream scientific outlet felt the need to ask the question so bluntly is itself the story — it signals that the earlier wave of confident claims about AI-driven drug discovery is now getting checked against the record, not just repeated.

That matters because most of what circulates about "AI drug discovery" online is marketing copy from the companies doing it, or hype-cycle commentary written without access to trial data. A piece that exists specifically to interrogate the claims — published by AnodicElegy on Science.org — is a different category of source: it's scrutiny, not promotion. If you're trying to form an accurate view of where the field stands, that's the source to start from, not a summary of a summary, and not this page either.

The detail

Here's the honest limit of what can be said from the record available for this piece: the Science.org article (So How Is AI Drug Discovery Doing, Really?) confirms that AI drug discovery is, as of August 2026, an active subject of scrutiny in the scientific press. It does not hand over a scoreboard here, and neither does this explainer — specific claims about approval rates, timelines, or named drug candidates aren't included below, because they aren't in the sourcing behind it. Anything more specific than "this is being actively debated in a serious publication" belongs in the original piece, not in a secondhand rewrite of it.

That's a deliberately narrow claim, and it's the right one to make. The AI-in-science space is full of writing that borrows the authority of "AI is transforming drug discovery" without citing anything that would let you check it. The Science.org piece is worth reading precisely because it's positioned to test that narrative rather than repeat it — and because a scrutiny piece from a peer-reviewed publisher carries a different weight than a press release recycled across a dozen blogs.

If you came here to evaluate the science itself, the move is simple: read the primary source before you form a view, and treat any secondary summary — including this one — as a pointer to it, not a substitute for it.

This explainer's remaining value is practical, not scientific. For readers who are also evaluating platforms to build AI-tools products on while they're at it, [check current price for Vercel Pro](Check current price for Vercel Pro) shows Vercel's current plan tier and what it includes — the kind of detail worth confirming directly rather than trusting a secondhand summary.

For readers who came here because they're building something in the AI-tools space — a research dashboard, an internal agent, a small product that sits on top of AI models — the rest of this page is for you.

What it means for you

If part of your work involves shipping AI-adjacent tools, the practical question isn't "is AI drug discovery working" — it's "what do I deploy this on." That's where this topic connects to a real decision rather than a hypothetical one.

Vercel is a hosting and deployment platform built for exactly that use case: shipping web apps and AI-powered tools without running your own servers. It publishes its plan tiers, including Pro and Enterprise, directly on its own pricing page rather than through third-party estimates, which matters because plan limits and features are the kind of detail that changes between updates and shouldn't be taken from a summary that's months old.

Before you commit to a plan, it's worth checking directly on the merchant's own page rather than a secondhand recap:

• Which plan tier your team size and traffic actually fall into.
• Whether the features you rely on today are included at that tier, or gated to the one above it.
• Whether Enterprise adds anything your current stack needs, or whether Pro already covers it.

Comparing how the current options actually stack up matters more than deciding on brand recognition alone — the Show HN Agentic Ship compared: which one is worth buying breakdown is a useful next stop if you're weighing more than one platform.

Best for: teams already building, or planning to build, AI-tools products — dashboards, agents, internal apps — who want deployment infrastructure without running their own servers.

Skip it if: you're not currently shipping a web-based tool. In that case, no hosting platform is the right purchase yet, Vercel included.

To see what a Pro or Enterprise plan actually includes today, rather than what a summary claims it includes, [see current pricing at the merchant](See current pricing at the merchant). That page is the only place this explainer treats as authoritative on cost, because it's the one place that gets updated when the numbers change.

The state of your own deployment stack is a question you can answer today, by checking what the platform you're considering actually offers.